Re: [Tracker] Full text search indexes only the first 64KiB of plaintext files?

[Aavo Tambur <aavo starman ee>]

2011/5/23 Martyn Russell <martyn lanedo com>:
> On 23/05/11 13:30, Aavo Tambur wrote:
> > The full text search of Tracker does not seem to index anything beyond
> > the first 64KiB of plaintext files.
>
> That value is not hard set upstream AFAIK. What version of Tracker are you
> using ?

I'm using Tracker distributed by the latest Ubuntu Natty, the 0.8.17.


> Depending on the version of Tracker, you can either change your .cfg file or
> GSettings. But there is an FTS config:
> ...
> $HOME/.config/tracker/tracker-extract.cfg:
>
> ÂMaxBytes=1048576

This setting did not exist... I downloaded the source package (from
Ubuntu) and found that that code did not have such parameter either.

As I'm a bit hesitant to compile the current version since it seems
that Ubuntu has made a number of patches on top of the code, I dug a
bit deeper into the 0.8.17 code.
I found the file [src/tracker-extract/tracker-extract-text.c] has two
hardcoded values: TEXT_MAX_SIZE at 1MiB and TEXT_CHECK_SIZE at
64KiB...
I changed both to be alot bigger, recompiled, repackaged and installed
and it seems to work... not fully sure though yet as re-indexing takes
a long time...

Thanks for nudging me to the right direction! :-)

Aavo